Induction and regulatory function of miR-9 in humanmonocytes and neutrophils exposed toproinflammatory signalsFlavia Bazzonia,1, Marzia Rossatoa, Marco Fabbrib, Daniele Gaudiosib, Massimiliano Mirolob, Laura Morib,Nicola Tamassiaa, Alberto Mantovanib, Marco A. Cassatellaa, and Massimo Locatib,2

aDepartment of Pathology, Division of General Pathology, University of Verona, Strada le Grazie 8, I-37134 Verona, Italy; and bLaboratory of LeukocyteBiology, Department of Translational Medicine, Istituto di Ricovero e Cura a Carattere Scientifico Istituto Clinico Humanitas, University of Milan, viaManzoni 56, I-20089 Rozzano, Italy

Edited by Bruce A. Beutler, The Scripps Research Institute, La Jolla, CA, and approved February 12, 2009 (received for review October 30, 2008)

Inflammation involves a coordinated, sequential, and self limitingsequence of events controlled by positive and negative regulatorymechanisms. Recent studies have shown that microRNAs (miRNAs),an evolutionarily conserved class of endogenous 22-nucleotidenoncoding RNAs, contribute to the regulation of inflammation byrepressing gene expression at the posttranscriptional level. In thisstudy, we characterize the profile of miRNAs induced by LPS inhuman polymorphonuclear neutrophils (PMN) and monocytes. Inparticular, we identify miR-9 as the only miRNA (among 365analyzed) up-regulated in both cell types after TLR4 activation.miR-9 is also induced by TLR2 and TLR7/8 agonists and by theproinflammatory cytokines TNF-� and IL-1�, but not by IFN�.Among the 3 different genes encoding miR-9 precursors in humans,we show that LPS selectively induces the transcription of miR-9–1located in the CROC4 locus, in a MyD88- and NF-�B-dependentmanner. In PMN and monocytes, LPS regulates NFKB1 at both thetranscriptional and posttranscriptional levels, and a conservedmiR-9 seed sustained a miR-9-dependent inhibition of the NFKB1transcript. Overall, these data suggest that TLR4-activated NF-�Brapidly increases the expression of miR-9 that operates a feedbackcontrol of the NF-�B-dependent responses by fine tuning theexpression of a key member of the NF-�B family.

inflammation � innate immunity � Toll-like receptors � cytokines � NFKB1

The innate immune response is the first line of defense againstinfectious agents and is mainly exerted by phagocytes, in-

cluding polymorphonuclear neutrophils (PMN) and monocyte/macrophages. This response is triggered by the recognition ofpathogen-associated molecular patterns of invading microorgan-isms by members of the Toll/IL-1 receptor (TLR) superfamily(1) among others. These receptors signal through similar intra-cellular pathways that start with the recruitment to the Toll-IL-1R (TIR) domain present in the receptor tail with 1 of 4possible TIR domain-containing adaptor molecules. The com-bination of adaptor molecules involved not only depends uponthe specific TLR engaged, but also defines the consequentcellular events. In particular, the MyD88 and TIR domain-containing adapter protein/MyD88 adapter-like protein(TIRAP/MAL) mediates the early NF-�B activation, while theTIR domain-containing adapter inducing IFN�/TIR-containingadapter molecule-1 (TRIF/TICAM-1) and TRIF-relatedadapter molecule/TIR-containing adapter molecule-2 (TRAM/TICAM-2) mediate the delayed NF-�B and IFN-regulatoryfactor (IRF) 3 signals (2, 3). As examples, TLR4 inducesproinflammatory cytokines via either a MyD88-dependent rapidactivation of the transcription factor NF-�B or costimulatory andantiviral proteins through a more delayed activation of bothNF-�B and IRF-3 mediated by TRIF. Conversely, TLR3 exclu-sively signals through the TRIF-dependent pathway and does notactivate the MyD88-dependent pathway (2, 3). Importantly, avariety of extracellular and intracellular negative feedback path-

ways have evolved to prevent an inappropriate inflammatoryresponse following activation of TLRs. These include the regu-lation of TLR expression, the production of molecules thatcompete with their ligand binding or signaling activities, and thegeneration of dominant negative splice variants or posttransla-tional modifications of signal transducers of the TLR signalingcascade (4, 5).

An emerging class of regulators of gene expression is repre-sented by microRNAs (miRNAs), which act at the posttran-scriptional level via an RNA interference mechanism (6, 7).miRNAs biogenesis involves the initial transcription by RNApolymerase II of primary miRNAs (pri-miRNA), which aresubsequently cleaved by the RNase III enzyme Drosha and Dicerto produce the 21- to 23-nt double-stranded RNA duplexes (6).The mature miRNA guide strand is then loaded into themiRNA-induced silencing complex, where it guides the recog-nition and translational repression or degradation of targetmRNAs (8). In mammals, a host of genes are processed toproduce over 700 miRNA (miRNA registry at www.sanger.ac.uk/Software/Rfam/mirna), which have been implicated in a widearray of biological processes ranging from cellular developmentand differentiation to tumors (6, 7). Recently, activation of theinnate immune response also has been associated with changesin the expression of selected miRNAs [namely miR-146 (9, 10),miR-155 (11, 12), miR-132 (10), and miR-125b (12)]. However,the ability of inflammatory ligands to modulate miRNA expres-sion and, more importantly, the role of regulated miRNAs in thedevelopment of an adequate immune response are just beginningto be explored.

Herein, we report the profiles of miRNAs induced by inflam-matory stimuli in human PMN and monocytes and identifymiR-9 as a previously unrecognized LPS-responsive miRNAinduced in a MyD88- and NF-�B-dependent manner in both celltypes. We also show that miR-9 takes part of a regulatorycircuitry controlling cell activation by means of inhibitory feed-back loop acting at the level of NFKB1, a transcriptionalregulator with a key role in the inflammatory response.

ResultsTo identify miRNAs potentially involved in the responses ofperipheral human PMN and monocytes to stimuli of bacterial

Author contributions: F.B., A.M., M.A.C., and M.L. designed research; M.R., D.G., M.M.,L.M., and N.T. performed research; M.F. analyzed data; and F.B. and M.L. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

1To whom correspondence may be addressed at: Department of Pathology, Division ofGeneral Pathology, University of Verona, Strada Le Grazie 4, I-37134 Verona, Italy. E-mail:flavia.bazzoni@univr.it.

2To whom correspondence may be addressed at: Laboratory of Leukocyte Biology, Istitutodi Ricovero e Cura a Carattere Scientifico Instituto Clinico Humanitas, University of Milan,Via Manzoni 113, I-20089 Rozzano, Italy. E-mail: massimo.locati@humanitas.it.

This article contains supporting information online at www.pnas.org/cgi/content/full/0810909106/DCSupplemental.

5282–5287 � PNAS � March 31, 2009 � vol. 106 � no. 13 www.pnas.org�cgi�doi�10.1073�pnas.0810909106

origin, the miRNA pattern of expression was investigated inPMN and monocytes stimulated for 8 h with 100 ng/ml LPS usinga TaqMan-based Low Density Array. As shown in Fig. 1, LPSinduced an up-regulation of 12 miRNAs in PMN and/or mono-cytes while no miRNA was significantly down-regulated (Fig. 1).LPS-induced miRNAs identified in the array were evaluated ina time-course analysis by RT-qPCR (Fig. 2). Consistent with thearray data, the expression of miR-155, miR-132, miR-146a,miR-146b, miR-187, miR-125a, miR-99b, and let-7e rapidlyincreased in LPS-treated monocytes but not in PMN (Fig. 2).Interestingly, both 3�-end (miR-9) and 5�-end (miR-9*) forms ofmiR-9 were the only miRNAs consistently induced by LPS inboth PMN and monocytes, being already detectable after 2 h andsteadily increasing over the time period assessed (Fig. 2 andsupporting information (SI) Fig. S1 A). Up-regulation of miR-9expression in PMN and monocytes stimulated with LPS for 24 hwas further confirmed in Northern blot analysis (Fig. S1B).Conversely, the induction of miR-222 and miR-196a observed inthe array analysis was not confirmed by RT-qPCR analysis (notshown). miR-9 was then chosen for a more detailed analysis,given that it is the only miRNA up-regulated in response to LPSin both cell types and that it has not been previously reported tobe involved in the inflammatory response.

LPS triggers different patterns of responses in PMN andmonocytes, partly because of the selective activation of the

different MyD88- and TRIF-dependent signaling pathwaysdownstream of the pattern-recognition receptor TLR4 (1). Toinvestigate the requirement of MyD88 and/or TRIF adaptors inthe induction of miR-9 expression by LPS and to evaluate miR-9regulation by other TLRs, PMN and monocytes were stimulatedwith Pam3CSK4 (100 ng/ml), a synthetic lipoprotein agonist atTLR2 that selectively requires TIRAP/MyD88; Resiquimod(R848, 10 �M), a TRL7/8 ligand signaling through MyD88 only;or polyinosinic:polycytidylic acid [poly(I:C)] (50 �g/ml), a syn-thetic mimetic of viral double-stranded RNA (dsRNA) thatinteracts with endosomal TLR3 and utilizes TRIF-mediatedsignaling (1). As shown in Fig. 3A, activation of TLR2 andTLR7/8 resulted in up-regulation of miR-9 expression in bothcell types, while that of TLR3 was ineffective. Conversely,poly(I:C) readily induced miR-155 in monocytes (Fig. S2), aspreviously reported in other cell types (11), demonstrating thatthe lack of miR-9 induction was not due to a general failure ofmonocytes to activate the TRIF-dependent pathway down-stream TLR3. In agreement with the lack of TLR3 expression inhuman PMN (13), poly(I:C) had no effect on miR-9 expressionin this cell type (Fig. 3A). Taken together, these data suggest thatin human phagocytes, activation of the MyD88-dependent sig-naling pathway is necessary and sufficient to increase miR-9expression in response to LPS and that no additional TRIF-dependent signals are required.

To test whether cytokines involved in the response to bacterialand/or viral infection are also effective at inducing miR-9expression, PMN and monocytes were stimulated with TNF-� (5ng/ml), IL-1� (20 ng/ml), IFN� (1000 U/ml), or IFN� (1000U/ml). The proinflammatory cytokines TNF-� and IL-1� in-creased miR-9 levels in both PMN and monocytes, while IFN�and IFN� were ineffective (Fig. 3A). Since miR-9 is up-regulated

Fig. 1. miRNAs induced by LPS in PMN and monocytes. PMN (A) andautologous monocytes (B) were cultured for 8 h in medium alone or in thepresence of 100 ng/ml LPS. The miRNA fraction was purified and changes inmiRNA expression levels were determined using a micro fluidic card as de-scribed in Materials and Methods. Results are expressed as arbitrary units ona log scale using RNU44 as reference control. The mean values of 2 individualexperiments performed are shown. Dotted lines represent the 2 and 0.5boundary values for fold induction.

Fig. 2. Kinetics of LPS-induced miRNAs in PMN and monocytes. PMN andmonocytes were cultured for the indicated times in medium alone (- -�- -PMN, - -�- - monocytes) or in the presence of 100 ng/ml LPS (—‘— PMN, —F—monocytes). miRNA fraction was purified and miR-9, miR-155, miR-132, miR-146a, miR-146b, miR-187, miR-125a, miR-99b, and let-7e expression was de-termined by RT-qPCR and normalized to the let-7a levels, as described inMaterials and Methods. The results are expressed as fold change and arerepresentative of 3 individual experiments.

Bazzoni et al. PNAS � March 31, 2009 � vol. 106 � no. 13 � 5283

IMM

UN

OLO

GY

by both LPS and TNF-�, we tested whether TLR4 induction ofmiR-9 required a TNF-� autocrine signaling as previouslyreported for miR-155 (11). Anti TNF� MoAbs completelyblocked miR-9 induction by TNF-�, but were ineffective whenLPS was used (Fig. 3B), indicating that TNF-� is not involved inthe induction of miR-9 by LPS. In addition, up-regulation ofmiR-9 expression was not modified by treatment with brefeldinA before LPS stimulation, ruling out the possible involvement ofsoluble mediators released in response to LPS for miR-9 up-regulation (Fig. 3B). Taken together, these data candidate miR-9as a novel miRNA involved in the responses of human phago-cytes to selected stimuli of bacterial origin or proinflammatorycytokines.

In both mouse and human genomes, miR-9 can be generatedby processing of 3 different miR-9 primary transcripts encodedby distinct genes (C1orf61 for pri-miR-9–1, BC036480 forpri-miR-9–2, and CR612213 for pri-miR-9–3, respectively). LPSinduced a time-dependent increase in pri-miR-9–1 levels andhad no effect on the other 2 miR-9 primary transcripts, both inPMN and monocytes (Fig. 4A). The miR-9–1 primary transcriptderives from the C1orf61 locus which encodes for CROC-4protein, a transcriptional activator for the c-fos proto-oncogene(14). An EST database analysis revealed the existence of aninternal product of the C1orf61 locus. Both transcriptional units(here called CROC-4a and CROC-4b: Fig. S3A) generate the

miR-9–1 precursor and are activated by LPS in PMN andmonocytes (Fig. S3B). Analysis of the C1orf61 locus with thetranscription start sites predictor SwitchGear software (availableat http://genome.ucsc.edu) supports the existence of an internaltranscriptional unit. Inspection of the genomic sequence located2 kb upstream of the predicted start sites of the 2 transcriptsidentified putative promoter regions with consensus bindingsites for known LPS-sensitive transcription factors, includingNF-�B (Fig. S3A). The observation that miR-9 induction by LPSdepends on the activation of the MyD88 pathway and theidentification of NF-�B consensus binding sites within the 2putative pri-miR-9–1 promoters suggested that the miR-9 in-duction by LPS may result from the transcriptional activity ofNF-�B. This hypothesis was confirmed by the suppressive effecton LPS-dependent miR-9 induction of NF-�B inhibitors (MG-132, BAY-117082 and PDTC) (Fig. 4B). In contrast, inhibitorsof p38 (SB-203580) and JNK (SP-600125) were ineffective (Fig.4B). Collectively, these data demonstrate that in PMN andmonocytes inflammatory stimuli sustain the NF-�B-dependenttransactivation of C1orf61 locus and consequent production ofpri-miR-9–1.

To gain insight on the biological relevance of miR-9 inductionunder inflammatory conditions, we searched for predictedmiR-9 targets, focusing our attention on regulators of transcrip-tion, which have been frequently shown to be preferentialmiRNA targets (15). In agreement with this, the public databaseof animal miRNA miRGen (available at http://www.diana.pcbi.upenn.edu/miRGen/v3/miRGen.html), which inte-grates analysis from PicTar (16), MiRanda (17), and TargetScan(18), predicted among high score miR-9 targets the transcrip-tional regulators Onecut2 and PRDM1/Blimp-1, which havebeen previously validated as miR-9 targets but are not expressedin PMN and monocytes (data not shown). A miR-9 seed was also

Fig. 3. miR-9 is induced by MyD88-activating TLR agonists and proinflam-matory cytokines. (A) PMN and monocytes were cultured for 8 h with 100ng/ml LPS, 100 ng/ml Pam3CSK4, 10 �M R848 or 50 �g/ml poly(I:C), TNF-�, 20ng/ml IL-1�, 1000 U/ml IFN�, or 1000 U/ml IFN�. miRNA fraction was extractedand analyzed for miR-9 expression by RT-qPCR. (B) monocytes and PMN werepretreated for 30 min with medium (black bars), 10 �g/ml anti-TNF-� MoAbs(hatched bars), or 5 �g/ml brefeldin A (gray bars) before stimulation withTNF-� or LPS. miRNA fraction was extracted after 8 h and analyzed for miR-9expression by RT-qPCR. miRNA expression is depicted as fold change unitsafter let-7a normalization. Data show one experiment representative of 3. nd:not determined.

Fig. 4. LPS up-regulates pri-miR-9–1 in a NF-�B-dependent manner. (A) PMNand monocytes were cultured in the presence or absence of LPS for theindicated time; total RNA was extracted and pri-miR-9–1 (—F—), pri-miR-9–2(—‘—), and pri-miR-9–3 (—�—) were analyzed by RT-qPCR and normalized tothe 18S RNA as described in Materials and Methods. Results show that onlypri-miR-9–1, but not pri-miR-9–2 or pri-miR-9–3, was induced by LPS. (B) PMNand monocytes were pretreated for 30 min with medium, 10 �M MG-132, 10�M BAY-117082, 300 �M PDTC, 20 �M SP-600125, or 10 �M SB-203580 andsubsequently cultured for 8 h with or without LPS. miR-9 expression levelswere determined by RT-qPCR and expressed as fold change after let-7anormalization. Data show 1 experiment representative of at least 2 for eachpanel.

5284 � www.pnas.org�cgi�doi�10.1073�pnas.0810909106 Bazzoni et al.

predicted in one of few highly conserved regions present in the3�-UTR of the NFKB1 gene (Fig. S4A). Interestingly, thepotential regulatory loop between NF-�B and miR-9 was alsoconfirmed by the miPromotor software. Thus, to test whethermiR-9 postranscriptionally affects NFKB1, reporter constructcontaining the renilla luciferase gene fused to the NFKB13�-UTR (luc-NFKB1) was transiently transfected in HEK-293cells together with expression plasmids encoding miR-9. Asshown in Fig. 5, miR-9 significantly reduced luc-NFKB1 lucif-erase activity, and the introduction of point mutations in themiR-9 seed in the 3�-UTR of NFKB1 (luc-mut-NFKB1) revertedthe inhibitory activity of miR-9, demonstrating that the NFKB13�-UTR contains an active seed of miR-9. The specificity ofmiR-9 was also demonstrated by the lack of effect on a c-fos3�-UTR reporter construct (luc-fos), which does not contain amiR-9 seed. Conversely, the c-fos 3�-UTR presents a miR-155seed (Fig. S4B) and was significantly inhibited by miR-155, inagreement with previous reports (19) (Fig. 5).

Finally, we analyzed the expression profile of NFKB1 in PMNand monocytes stimulated with LPS at the transcript and proteinlevels (Fig. 6). Interestingly, LPS rapidly induced a consistentincrease in NFKB1 transcripts in PMN and monocytes (Fig. 6A),but the transcript up-regulation was not paralleled by a compa-rable increase of the corresponding protein, which showedconstant expression levels (Fig. 6B). These results indicate thatNFKB1 is subjected to a LPS-dependent regulation at thetranscriptional level and suggest that a second regulatory mech-anism, acting at the posttranscriptional level, is also operative. Inagreement with this finding, monocytes overexpressing miR-9(Fig. 6C) showed a reduced expression of the endogenousNFKB1/p105 protein (Fig. 6D). Conversely, transfection ofmonocytes with either the pcDNA3 empty vector or with themiR-155-encoding vector (Fig. 6C) did not alter the levels ofexpression of the endogenous NFKB1/p105 protein (Fig. 6D),confirming the specificity of the effect of miR-9.

DiscussionIn the last few years, miRNAs have emerged as a class of geneexpression regulators involved in a variety of biological processes(7, 20), and recent evidence indicates that they also play a majorrole in immunity (21). In particular, miRNAs have been involvedin myeloid and lymphoid differentiation (22–25), as well as in theinduction of an appropriate adaptive immune response (25–27).

A rapid increase in the expression of selected miRNAs has alsobeen recently observed during the activation of an innateimmune response (10–12), but the relevance of miRNAs in thisprocess is just beginning to be explored.

In this study, we investigated the potential involvement ofmiRNAs in the regulation of innate immune response by eval-uating the profile of miRNAs expressed in human PMN andmonocytes after TLR engagement or cytokine activation. AnRT-qPCR-based screening on human monocytes exposed toLPS not only confirmed the induction of miRNAs previouslydescribed in monocytic cell lines or mouse macrophages, namelyof miR-146 (10), miR-155 (11, 12), and miR-132 (10), but also

Fig. 5. The NFKB1 gene is a molecular target of miR-9. The indicatedluciferase constructs (luc vectors) were cotransfected with expression vectorsencoding miR-155 (gray columns), miR-9 (black columns), or the pcDNA3empty vector (white columns). Results are expressed as mean (� SD, n � 3) ofthe ratio between renilla luciferase and firefly control luciferase activities(RLU), adjusted to 1. **: P � 0.01; ns: P � 0.05.

Fig. 6. Analysis of the endogenous miR-9 target: NFKB1 mRNA and proteinexpression. (A) PMN and monocytes were cultured in the presence of LPS forthe indicated times. Total RNA was purified and used to assay NFKB1 mRNAexpression by RT-qPCR, as described in Materials and Methods. Relative NFKB1gene expression is depicted as MNE units after GAPDH normalization. Datareported are representative of 3 independent experiments. (B) Whole-cellextracts (20 �g) were usually loaded on gels and immunoblots were per-formed by simultaneously using Abs specific for NFKB1 and Abs specific foractin, followed by incubation with Alexa Fluor-680 goat anti-rabbit Abs. Oneexperiment representative of 3 is shown. The relative NFKB1/p105 levels (� SD,n � 3), quantified as described in SI Materials and Methods, are reportedbelow each panel. (C) 6 � 106 monocytes were transfected with 5 �g ofpcDNA3 empty vector, miR-155-encoding vector, or miR-9-encoding vector asdescribed in Material and Methods. 48 hours posttransfection total RNA waspurified and analyzed for miR-9 expression by RT-qPCR. Relative miR-9 ex-pression is represented as MNE units after let-7a normalization. (D) Fifteenmicrograms of whole cell extracts, purified from transfected monocytes cul-tured in the same conditions as in (C), were usually loaded on gels andNFKB1/p105 protein was detected as described above. The relative NFKB1/p105 levels, normalized for the total actin, are reported below the Westernblot. (C and D) One experiment representative of 2.

Bazzoni et al. PNAS � March 31, 2009 � vol. 106 � no. 13 � 5285

IMM

UN

OLO

GY

uncovered the induction of previously unrecognized miRNAs,including miR-187, the miR-125a/miR-99b/let-7e cluster, andmiR-9/9*. Interestingly, out of the 365 tested, miR-9/9* was theonly miRNA induced by LPS also in human PMN.

Initially identified as a brain-specific miRNA, miR-9 has beenimplicated in mammalian neuronal development and function,notably by inhibiting expression of antineurogenic transcriptionfactor-encoding genes (28, 29). Outside the central nervoussystem, miR-9 was subsequently found to play a critical role inthe control of the secretory function of insulin-producing cells bymaintaining appropriate levels of the transcription factor One-cut2 (30), and more recently it was shown to down-regulate thetranscription factor PRDM1/Blimp-1 in Hodgkin/Reed-Sternberg cells, thus interfering with normal B-cell terminaldifferentiation and contributing to the pathogenesis of Hodgkinlymphoma (31). To our knowledge, the identification of miR-9as an LPS-responsive miRNA in primary human PMN andmonocytes represents the first evidence linking miR-9 to theinnate immune response.

Several lines of evidence support the notion that miR-9 isdirectly induced by LPS via the MyD88/NF-�B-dependent path-way. First, LPS efficiently up-regulated miR-9 expression inhuman PMN, in which TRL4 engagement does not activate theTRIF-dependent pathway (32). Second, activation of TLR3, theonly TLR known to transduce its signal independently fromMyD88 (1), does not modulate miR-9 expression. This obser-vation supports the conclusion that in monocytes the inductionof miR-9 proceeds in a MyD88-dependent manner, even thoughin these cells both MyD88- and TRIF-dependent signalingpathways can be simultaneously activated downstream TLR4. Bycontrast, activation of TRIF-dependent pathway downstream ofTLR3 is sufficient to induce miR-155 expression in monocytes aspreviously reported (11). Finally, activation of TLR2 andTLR7/8, known to engage only the MyD88-dependent pathway,led to increased miR-9 expression in both PMN and monocytes.Taken together, these data rule out any contribution of theTRIF-dependent pathway downstream of TLR4 to LPS-inducedmiR-9 expression. TLR4 activation also leads to the productionof proinflammatory cytokines, such as TNF� and IL-1�, in bothPMN and monocytes and to the production of antiviral type IIFN in monocytes. In turn, these mediators can signal throughtheir receptor in an autocrine fashion. However, we providedemonstrations indicating that induction of miR-9 by LPS occursin a direct manner, without the involvement of endogenousmediators. First, activation of PMN and monocytes with IFN�does not promote miR-9 expression. Secondly, even thoughTNF� is able to induce miR-9 expression, its presence is notrequired for TLR4-mediated production of miR-9 because in-cubation of PMN and monocytes with anti-TNF� antibodiesbefore LPS stimulation did not suppress the up-regulation ofmiR-9 expression. Finally, inhibition of protein secretion bybrefeldin A does not impair LPS-induced miR-9 up-regulation,ruling out the involvement of any LPS-induced endogenousmediator.

In the human genome, 3 distinct genes encode differentpri-miR-9 leading to the generation of identical mature miR-9.Notwithstanding, quantitative analysis of the levels of expressionof each pri-miR-9 demonstrated that miR-9 accumulation inresponse to LPS entirely derived from the pri-miR-9–1. Adetailed analysis of EST database revealed the existence in theC1orf61 locus of 2 transcriptional units regulated by independentpromoters. Both transcriptional units were activated by LPS, andinspection of the upstream promoter regions of miR-9–1, to-gether with the analysis of miR-9 induction in the presence ofdifferent signaling pathway inhibitors, showed that NF-�B acti-vation is involved in the induction of this miRNA.

One model for miRNA-based regulation of gene expressionenvisages a global effect on the cell transcriptome mediated by

a direct effect of the miRNA on a large number of targettranscripts, which are diminished to negligible levels. In agree-ment with this model, an overrepresentation of the seeds ofmiRNAs induced in a defined experimental condition in the3�-UTR of genes down-regulated in the same condition has beendemonstrated, as in the case of the miR-155 seed in LPS-down-regulated genes (26). An alternative mechanism, also sustainedby the frequent observation of a complex cross-talk betweenmiRNAs and transcription factors (19, 22, 24, 26), foresees anindirect effect of miRNAs on the cell transcriptome mediated bythe control of transcripts encoding for key transcription regu-lators. As this mechanism has also been previously reported formiR-9 (30, 31), among predicted miR-9 targets we focused ourattention on genes directly involved in the control of transcrip-tion under inflammatory conditions. Cotransfection experi-ments of miR-9 with 3�UTR luciferase reporters identifiedNFKB1 as miR-9 target. Interestingly, the nfkb1 gene encodes 2functional proteins (p50 and its precursor p105) with distinctbiological activities during LPS responses (33). The active nfkb1gene product (p50) can form heterodimers with other NF-�Bsubunits, which function as transcriptional activators. Alterna-tively, p50 can form homodimers that have been shown to fulfillan anti-inflammatory role by attenuating transcription of proin-flammatory cytokines and by activating IL-10 expression (34,35). Notably, the relative amount of each of the NF-�B familymembers may affect the outcome of the innate immune response(34, 35). Unbalanced expression of p50 has been observed underpathological conditions in which negative pathways of regulationprevail such as LPS tolerance, chronic inflammatory conditions,and cancer (36, 37). Analysis of the endogenous levels of thismiR-9 target showed that NFKB1 mRNA rapidly and transientlyincreased during the initial hours of the LPS response, while theNFKB1/p105 protein levels remained constant for up to 24 h.This behavior perfectly fits the ‘‘micromanaging model,’’ ac-cording to which miRNAs contribute to maintain the optimallevel of expression of some genes, particularly regulatory genes,that might have a narrow window of optimal expression (20).Consistent with this hypothesis, we found that miR-9-overexpressing monocytes display decreased NFKB1/p105 lev-els. Conversely, variations in NFKB1/p105 expression were notobserved after transfecting miR-155, indicating that this effect isspecific for miR-9. Since NF-�B is a key regulator of inflam-mation, the NFKB1 levels are likely to be a strictly controlled andtimely regulated event of relevance for the proper progression ofthe inflammatory response. On the basis of our observations itis tempting to propose a model in which the parallel NF-�B-dependent induction of NFKB1 and miR-9 provides a mean tosmooth out the fluctuations in gene expression and fine tune thesynthesis of this key transcription factor, thus allowing theproinflammatory phase of the LPS response to correctly pro-ceed. It will be important to assess how pathological conditionsin which regulatory circuits of inflammation prevail, such as thesystemic anti-inflammatory response syndrome associated withLPS tolerance and cancer, affect the miR-9 regulatory axis.

In summary, the results reported here extend to freshlyisolated human monocytes and PMN previous observations onthe induction of miRNAs by LPS in mouse macrophages and celllines. In addition to the classic miR-155 and miR-146, a new setof miRNAs (miR-9, miR-187, miR-125a, miR-99b, and let-7e)were found to be LPS-responsive miRNAs in human monocytes.Among these, only miR-9 was also induced in neutrophils.Induction of miR-9 was also mediated by the proinflammatorycytokines IL-1� and TNF-�, but not by IFN�. Interestingly,NFKB1/p105/p50 was identified in silico and experimentally asa miR-9 target. Given the regulatory function of p50 ho-modimers these results suggest a model whereby induction ofmiR-9 acts a tuning mechanism to prevent negative regulation by

5286 � www.pnas.org�cgi�doi�10.1073�pnas.0810909106 Bazzoni et al.

p50 homodimers as occurs in monocytes in systemic anti-inflammatory response syndrome and in cancer.

Materials and MethodsMaterials. The detailed list of the materials is provided in SI Materials andMethods.

Cell Purification and Culture. Human PMN and monocytes were purified andcultured as described in SI Materials and Methods. The HEK293 cell line wasgrown in DMEM (Cambrex) supplemented with 10% FCS (Euroclone), 100 U/mlpenicillin/streptomycin (Cambrex), and 2 mM L-glutammine (Cambrex).

Quantification of miRNAs Expression Level. PMN and monocytes were stimu-lated with 100 ng/ml LPS for 8 h and the RNA fraction that is highly enrichedfor small RNA species (� 200 bp) was isolated by using the mirVana isolationkit (Ambion, Applied Biosystems), according to the manufacturer’s protocol.The small RNA fractions were reverse transcribed and the analysis of theexpression level of 365 miRNA was performed using a TaqMan-based LowDensity Array. Details on reverse transcription as well as TaqMan Array arepresented in SI Materials and Methods. Experimental data were then ana-lyzed by SDS2.2.2 software and the relative miRNA expression values werecalculated using RNU44 as endogenous control (38). miRNAs with a thresholdcycle �33 that showed a fold change �2 or �0.5 in samples treated with LPSas compared to control samples were considered as differentially expressed.

Real Time RT-PCR (RT-qPCR). miRNAs differentially expressed were validatedusing individual TaqMan miRNA Assay (Applied Biosystems) as described indetails in SI Materials and Methods. The expression of miR-9 precursors(pri-miR-9–1, pri-miR-9–2 and pri-miR-9–3), CROC-4a, CROC-4b and NFKB1was quantified by RT-qPCR (see SI Materials and Methods for details). Theprimers and probes used are described in Table S1.

Northern Blot. Small RNA fraction was purified from PMN and monocytes andprocessed for Northern blot analysis as described (39) and detailed in SIMaterials and Methods.

Constructs Generation and Luciferase Reporter Assay. Generation of miR-9-,miR-155-encoding vectors, luc-fos, luc-NFKB1 and luc-mut-NFKB1 togetherwith the luciferase reporter assay are described in SI Materials and Methods.The oligonucleotides used to generate the constructs are listed in Table S1.

Monocytes Transfection. Freshly purified monocytes (6 � 106) were transfectedwith 5 �g of plasmid DNA (pcDNA3 empty vector, pcDNA3-miR-155, orpcDNA3-miR-9) using the Amaxa Nuclofector and the Human MonocyteNucleofector kit (Amaxa), according to the manufacturer’s protocol. Cellswere then cultured and processed as detailed in SI Materials and Methods.

Western Blot. Preparation of cell lysates and Western blot analysis wereconducted as previously described (40) and described in details in SI Materialsand Methods.

Statistical Analysis. Data are expressed as means � SD. Statistical changes inluciferase expression were determined using the one-way ANOVA with � setto 0.05 according to the Newman-Keuls test.

ACKNOWLEDGMENTS. This work was supported by grants from Ministerodell’Istruzione, dell’Universita e della Ricerca (PRIN and FIRB projects), theFondazione Cariverona, the University of Verona (Joint Project grant), theFondazione Cariplo (NOBEL project), and the Programma Straordinario diRicerca Oncologica 2006 from Alleanza Contro il Cancro and Istituto Superioredi Sanita (RNBIO project). The experimental work was conducted in part in thecontext and with the support of the Fondazione Humanitas per la Ricerca(Rozzano, Milan, Italy). The generous contribution of the Italian Associationfor Cancer Research is gratefully acknowledged.

1. Akira S, Uematsu S, Takeuchi O (2006) Pathogen recognition and innate immunity. Cell124:783–801.

2. Akira S, Takeda K (2004) Toll-like receptor signalling. Nat Rev Immunol 4:499–511.3. O’Neill LA (2006) How Toll-like receptors signal: What we know and what we don’t

know. Curr Opin Immunol 18:3–9.4. Liew FY, Xu D, Brint EK, O’Neill LA (2005) Negative regulation of toll-like receptor-

mediated immune responses. Nat Rev Immunol 5:446–458.5. Mantovani A, Locati M, Polentarutti N, Vecchi A, Garlanda C (2004) Extracellular and

intracellular decoys in the tuning of inflammatory cytokines and Toll-like receptors:The new entry TIR8/SIGIRR. J Leukoc Biol 75:738–742.

6. Bartel DP (2004) MicroRNAs: Genomics, biogenesis, mechanism, and function. Cell116:281–297.

7. Ambros V (2004) The functions of animal microRNAs. Nature 431:350–355.8. Liu J (2008) Control of protein synthesis and mRNA degradation by microRNAs. Curr

Opin Cell Biol 20:214–221.9. Perry MM, et al. (2008) Rapid changes in microRNA-146a expression negatively regu-

late the IL-1beta-induced inflammatory response in human lung alveolar epithelialcells. J Immunol 180:5689–5698.

10. Taganov KD, Boldin MP, Chang KJ, Baltimore D (2006) NF-kappaB-dependent induc-tion of microRNA miR-146, an inhibitor targeted to signaling proteins of innateimmune responses. Proc Natl Acad Sci USA 103:12481–12486.

11. O’Connell RM, Taganov KD, Boldin MP, Cheng G, Baltimore D (2007) MicroRNA-155 isinduced during the macrophage inflammatory response. Proc Natl Acad Sci USA104:1604–1609.

12. Tili E, et al. (2007) Modulation of miR-155 and miR-125b levels following lipopolysac-charide/TNF-alpha stimulation and their possible roles in regulating the response toendotoxin shock. J Immunol 179:5082–5089.

13. Tamassia N, et al. (2008) Activation of an immunoregulatory and antiviral geneexpression program in poly(I:C)-transfected human neutrophils. J Immunol 181:6563–6573.

14. Jeffrey PL, et al. (2000) CROC-4: A novel brain specific transcriptional activator of c-fosexpressed from proliferation through to maturation of multiple neuronal cell types.Mol Cell Neurosci 16:185–196.

15. Asirvatham AJ, Gregorie CJ, Hu Z, Magner WJ, Tomasi TB (2008) MicroRNA targets inimmune genes and the Dicer/Argonaute and ARE machinery components. Mol Immu-nol 45:1995–2006.

16. Krek A, et al. (2005) Combinatorial microRNA target predictions. Nat Genet 37:495–500.

17. John B, et al. (2004) Human MicroRNA targets. PLoS Biol 2:e363.18. Lewis BP, Shih IH, Jones-Rhoades MW, Bartel DP, Burge CB (2003) Prediction of

mammalian microRNA targets. Cell 115:787–798.19. Gottwein EN, et al. (2007) A viral microRNA functions as an orthologue of cellular

miR-155. Nature 450:1096–1099.20. Bartel DP, Chen CZ (2004) Micromanagers of gene expression: The potentially wide-

spread influence of metazoan microRNAs. Nat Rev Genet 5:396–400.

21. Taganov KD, Boldin MP, Baltimore D (2007) MicroRNAs and immunity: Tiny players ina big field. Immunity 26:133–137.

22. Fazi F, et al. (2005) A minicircuitry comprised of microRNA-223 and transcription factorsNFI-A and C/EBPalpha regulates human granulopoiesis. Cell 123:819–831.

23. Fontana L, et al. (2007) MicroRNAs 17–5p-20a–106a control monocytopoiesis throughAML1 targeting and M-CSF receptor upregulation. Nat Cell Biol 9:775–787.

24. Xiao C, et al. (2007) MiR-150 controls B cell differentiation by targeting the transcrip-tion factor c-Myb. Cell 131:146–159.

25. Cobb BS, et al. (2005) T cell lineage choice and differentiation in the absence of theRNase III enzyme Dicer. J Exp Med 201:1367–1373.

26. Rodriguez A, et al. (2007) Requirement of bic/microRNA-155 for normal immunefunction. Science 316:608–611.

27. Thai TH, et al. (2007) Regulation of the germinal center response by microRNA-155.Science 316:604–608.

28. Sempere LF, et al. (2004) Expression profiling of mammalian microRNAs uncovers asubset of brain-expressed microRNAs with possible roles in murine and human neu-ronal differentiation. Genome Biol 5:R13.

29. Leucht C, et al. (2008) MicroRNA-9 directs late organizer activity of the midbrain-hindbrain boundary. Nat Neurosci 11:641–648.

30. Plaisance V, et al. (2006) MicroRNA-9 controls the expression of Granuphilin/Slp4 andthe secretory response of insulin-producing cells. J Biol Chem 281:26932–26942.

31. Nie K, et al. (2008) MicroRNA-mediated down-regulation of PRDM1/Blimp-1 inHodgkin/Reed-Sternberg cells: A potential pathogenetic lesion in Hodgkin lympho-mas. Am J Pathol 173:242–252.

32. Tamassia N, et al. (2007) The MyD88-independent pathway is not mobilized in humanneutrophils stimulated via TLR4. J Immunol 178:7344–7356.

33. Pereira SG, Oakley F (2008) Nuclear factor-kappaB1: Regulation and function. IntJ Biochem Cell Biol 40:1425–1430.

34. Cao S, Zhang X, Edwards JP, Mosser DM (2006) NF-kappaB1 (p50) homodimers differ-entially regulate pro- and anti-inflammatory cytokines in macrophages. J Biol Chem281:26041–26050.

35. Ghosh S, May MJ, Kopp EB (1998) NF-kappa B and Rel proteins: Evolutionarily con-served mediators of immune responses. Annu Rev Immunol 16:225–260.

36. Karin M (2006) Nuclear factor-kappaB in cancer development and progression. Nature441:431–436.

37. Saccani A, et al. (2006) p50 nuclear factor-kappaB overexpression in tumor-associatedmacrophages inhibits M1 inflammatory responses and antitumor resistance. CancerRes 66:11432–11440.

38. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-timequantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25:402–408.

39. Varallyay E, Burgyan J, Havelda Z (2008) MicroRNA detection by northern blottingusing locked nucleic acid probes. Nat Protoc 3:190–196.

40. Crepaldi L, et al. (2001) Up-regulation of IL-10R1 expression is required to renderhuman neutrophils fully responsive to IL-10. J Immunol 167:2312–2322.

Bazzoni et al. PNAS � March 31, 2009 � vol. 106 � no. 13 � 5287

IMM

UN

OLO

GY